Thread by @PhDniX

If you open up modern grammars of Arabic today you will learn that the li- of command is syncopated when wa- or fa- precede. I.e. li-yaktub "let him write" but fa-l-yaktub "so let him write".

But was this always the rule for Classical Arabic? Doesn't seem to be the case! THREAD

When reading Sībawayh, he clearly notes that those who retain the vowels in wa-huwa "and he" and fa-hiya "so she" (instead of wa-hwa and fa-hya), also retain the i-vowel in wa-li-yaktub rather than the now considered Classical wa-l-yaktub.

That's quite different from how Classical Arabic grammar is viewed today, there wa-l-yaktub forms are regularly syncopated, while wa-huwa and fa-hiya as a rule are not.

Is Sībawayh (Baṣran) just confused or is this perhaps a difference of opinion with the Kufan grammarians?

Al-Farrāʾ (Kufan) seems to imply much the same thing, explicitly connecting the loss of vowels in fa-hwa, fa-hya, la-hwa but also harmun (<harumun) and rajlun (rajulun) with the dropping of the vowel in the li- of command.

So this isn't a difference in grammatical school either!

Next I had a look at al-ʾAḫfaš, Sībawayh's only significant student. His wording frequently shows he does not take the form without the vowel as the default: "as for <wa-l(i)-takun> (Q3:104), some of them drop the vowel of the lām as well."

Traces that the li- of command could occur unsyncopated has actually come down to us through the Qirāʾāt tradition as well. The pseudo-canonical reader al-Ḥasan al-Baṣrī (one of the 4 after the 10) frequently, though without any rhyme or reason, reads li- in such positions.

He reads:
Q2:282 wa-li-yumlil, wa-li-yattaq (but fa-l-yaktub and fa-l-yumlil in the same verse!)
Q4:9 wa-li-yaḫša, wa-li-yattaqū, wa-li-yaqūlū
Q4:102 fa-li-taqum (but not wa-li-yaʾḫuḏū, etc. in the same verse)
Q10:58 fa-li-tafraḥū

So not only do early grammarians seem to allow for unsyncopated forms (and explicitly endorse it if one says wa-huwa and fa-hiya as well), we even find traces of this in the (non-canonical) readings.

This clearly suggests that this Classical norm had simply not yet taken form.

There's at least one case attributed to a canonical reader, ʾAbū ʿAmr (through a now non-canonical transmitter: ʿAbbās b. al-Faḍl): Q24:31 wa-li-yaḍribna. Ibn Mujāhid (d. 324) is confused: "I don't know what this is"
So by his time, the classical norm seems to have established.

This is because to Ibn Mujāhid not dropping the vowel in this case causes it to be a "li- with the meaning of kay 'in order to'", which makes little sense here. That is the way it works in today's Classical Arabic as well, but clearly no in Sībawayh's (or al-Ḥasan's) time.

In Kufic vocalised manuscripts, unsyncopated li-s of command occur frequently:
Q2:282 fa-li-yumlil [secondary reading] Arabe 351
Q4:9 wa-li-yaḫša Arabe 339
Q4:102 wa-li-taqum Arabe 330f
Q24:31 wa-li-yaḍribna Arabe 347a [secondary reading]

I ran into these with two manuscripts I've personally worked on as well.
Arabe 334a has Q5:11 fa-li-tawakkali, which at the time I called "perhaps ungrammatical", but is clearly commonplace.
Arabe 329f Q10:58 fa-li-yafraḥū (green dot marks the secondary fa-l-yafraḥū!)

What's striking is that none of these manuscripts I've mentioned here seem to represent al-Ḥasan's reading; Certainly not the last two, which are both non-canonical readings that are very distinct from his reading. Clearly such forms were much more widespread than the are now.

This detail, of course, has zero effect on how we understand the text, but has big implications as to how we understand Classical Arabic grammar. Too often we project back our modern textbooks to the early and even pre-Islamic period. The grammar of Classical Arabic is a process.

We can see grammarians negotiating different norms, to eventually end up at the unified Classical Arabic that forms the basis for Modern Standard Arabic. Vocalised Qurans, but also simply Quranic readings serve as evidence that what grammarians discuss actually reflect a reality.